Crosstalk between bone metastatic cancer cells and sensory nerves in bone metastatic progression

Study approval

All human studies and all animal studies followed the Declaration of Helsinki and the Institutional Animal Care and Use Committee Guidelines, respectively. All animal studies were approved by the Institutional Animal Care and Use Committee at Wake Forest University School of Medicine (Protocol A15-056, A17-047, A18-026, A19-173, A19-192, A21-029). All human studies performed were approved by either (i) the Institutional Review Board at Jikei University School of Medicine (# 28-140[8383], 30-136[9157]; IRB) or (ii) the IRB at Fred Hutchinson Cancer Center (# IR6312 and IR7917; IRB). Informed consent was obtained from all subjects involved in the current study.

Cell culture

Human prostate cancer cell lines PC-3 (Cat. #: CRL-1435), DU145 (Cat. #: HTB-81), and LNCaP (Cat. #: CRL-1740); human breast cancer cell lines MDA MB-231 (Cat. #: HTB-26), MCF-7 (Cat. #: HTB-22), and ZR75-1 (Cat. #: CRL-1500); human lung cancer cell lines A549 (Cat. #: CCL-185) and SK-MES-1 (Cat. #: HTB-58); murine prostate cancer cell line RM-1 (Cat. #: CRL-3310); and human lung epithelial cell line BEAS-2B (Cat. #: CRL-9609) were purchased from the American Type Culture Collection (ATCC). PC-3, DU145, LNCaP, ZR75-1, and A549 cells were maintained in Roswell Park Memorial Institute 1640 medium (Cat. #: 11-875-093; Thermo Fisher Scientific [Gibco]). MDA-MB-231, MCF7, SK-MES-1, and RM-1 cells were maintained in DMEM (Cat. #: 11-995-073; Thermo Fisher Scientific [Gibco]). BEAS-2B cells were maintained in BEGM (Cat. #: CC-3170; Lonza). All cultures were supplemented with 10% (V/V) FBS (Cat. #: 26-140-079; Thermo Fisher Scientific [Gibco]), 1% (V/V) penicillin–streptomycin (Cat. #: 15-140-163; Thermo Fisher Scientific [Gibco]), and 1% (V/V) L-glutamine (Cat. #: 25-030-164; Thermo Fisher Scientific [Gibco]). Cells were incubated at 37°C, 5% CO₂, and 100% humidity, and were routinely passaged when they were no more than 80% confluent.

Before the animal studies, some cancer cell lines (PC-3, DU145, A549, and RM-1) were transformed to stably express GFP and firefly luciferase by transduction with a lentivirus (Lenti-GF1-CMV-VSVG) generated by the University of Michigan Vector Core (Eber et al, 2021). The transduced cells were sorted for GFP-positive cells at the Wake Forest Baptist Comprehensive Cancer Center Flow Cytometry Shared Resource using Astrios EQ (Beckman Coulter), expanded, and frozen at low passage (<10).

Conditioned medium collection

To collect the conditioned medium (CM) from control or cancer cells, 5 × 10⁵ cells were seeded onto 10-cm dishes in complete growth medium. After 24 h, the medium was replaced with 10 ml of serum-free corresponding growth medium. For the control CM, 10 ml of serum-free DMEM or Roswell Park Memorial Institute medium was added to a 10-cm dish without adding any cancer cells. After 24 h of incubation at 37°C with 5% CO₂, the CM was collected, filtered through a 0.2-μm syringe filter (Cat. #. CLS431222; MilliporeSigma) to remove any cell debris, and then stored at 4°C until use.

Intrafemoral injection mouse model

Luciferase-expressing cancer cells were inoculated intrafemorally into mice. We used this previously well-established approach to establish tumor within the marrow (Schwei et al, 1999). Briefly, a one-cm incision was made in the skin of the right hindlimb (lateral side parallel to the femur) of mice to expose the muscle. The rectus femoris and vastus medialis muscles were separated using the line of connective tissue as a guide. Thereafter, the rectus femoris muscle and patella were moved to the medial side of the knee to expose the condyles of the femur. The connective tissue between the femur and the patella was cut, but the patellar tendon was not. A 27G needle was used to create the hole on the femur and then replaced with a C313I injector (Cat. #: C313I; Plastics One, Inc.) with Tygon tubing. After confirming that the injector was in the intramedullary space by an X-ray (Faxitron Bioptics MultiFocus X-ray System), cancer cells (suspended in 5–10 μl of Hanks’ buffered saline solution [Cat. #: 14175103; Thermo Fisher Scientific (Gibco)]) were injected using a 10-μl Hamilton syringe (Cat. #: 80300; Hamilton Company). The injection site was plugged with bone cement to delay the spread of the tumor into the adjacent soft tissue. The patella was then gently returned to its correct orientation, and to avoid possible patella displacement, muscles were secured back into position using a horizontal mattress suture technique and 7–0 absorbable sutures (Cat. #: 07-809-2011; Patterson Veterinary). Wound closure was then achieved with absorbable sutures. The same surgical procedure was used for sham animals except that the same amount of Hanks’ buffered saline solution was injected instead of the cancer cells. Thereafter, tumor growth, CIBP behaviors, and bone remodeling were measured, as described below.

For the xenograft model, PC3, DU145, or A549 cells (4 × 10⁴ cells/10 μl) were inoculated into SCID Hairless Outbred mice (male, 4–6 wk old, Crl:SHO-Prkdc^scidHr^hr, Cat. #: 474; Charles River Laboratories). For the syngeneic model, RM-1 cells (1 × 10³ cells/5 μl) were inoculated into CGRP homozygous (CGRP KO) (male, 4–6 wk old, B6.129P2[Cg]-Calca^{tm1.1(EGFP/HBEGF)Mjz}/Mmnc, Cat. #: 036773-UNC; Mutant Mouse Resource and Research Center at Jackson Laboratory), CGRP heterozygous (Control), or C57BL/6J (male, 4–6 wk old, Cat. #: 000664; Jackson Laboratory) mice. In some cases, mice were treated with either mouse anti-CGRP monoclonal antibody or isotype control antibody (30 mg/kg, intraperitoneal injection; TEVA Pharmaceuticals) or vehicle or CGRP8-37 (100 µg/kg, intraperitoneal injection; R&D Systems).

Measures of tumor growth

Bioluminescent images were obtained using IVIS Lumina LT Series III (PerkinElmer) through the Cell and Viral Vector Core Laboratory of the Wake Forest Baptist Comprehensive Cancer Center. Briefly, luciferase-expressing cancer cell–bearing mice were injected with luciferin (30 mg/ml, Cat. #: 122799; PerkinElmer) by intraperitoneal injections and ventral images were acquired 12 min post-injection under 2.5% isoflurane/air anesthesia. The total tumor burden of each animal was calculated using regions of interest (ROI) that encompassed the entire tumor area.

Measures of CIBP behavior

Guarding behavior measurements were performed twice a week as previously described (Peters et al, 2004; Jimenez-Andrade et al, 2011). Tumor-bearing mice demonstrate a progressive increase in spontaneous pain behaviors, including guarding and periodic flinching of the inoculated limb. Mice were placed on nylon mesh platforms in clear plastic enclosures and acclimated for 30 min. Mice were observed for 5 min by an experimenter blinded to treatment groups.

The running wheel assay was performed five times a week using commercially available equipment (Med Associates, Inc.). Standard running wheel chambers contained a running wheel (18.54 cm diameter, 58.2 cm circumference) with a stainless-steel grid bar running surface located outside a 27.15 × 20.8 × 15.39 cm polycarbonate cage. Animals had free access to running wheels through a side opening in the cage. Each quarter turn of the running wheel operated a microswitch, and closures were recorded via a PC-compatible interface and computer (Med Associates, Inc.). Total microswitch closures (responses) and the time elapsed between microswitch closures (inter-response time [IRT]) were recorded using the MED-PC programming language. Each running wheel chamber was isolated in a PVC sound- and light-attenuating enclosure with a ventilator fan. For experiments involving running wheel assessment, mice were housed under a reverse light:dark cycle and running wheel sessions were conducted during the dark phase of the light:dark cycle on weekdays only. Mice were allowed free access individually to running wheels for 30-min sessions. Baseline measures for distance and optimal running rate were obtained by averaging data from the last five sessions of this 3-wk period. In order to easily condition IRT data for distance traveled in running wheels and optimal running rate analysis, a Java program was developed to truncate and transfer the data daily from the MED-PC programming language into text files that would be easily importable into SAS. Responses were converted to distance traveled (14.55 cm per response), and IRTs were converted to speed (14.55 cm/IRT in seconds, cm/s). For optimal velocity determination, a paradigm was developed to find the running rates that best distinguished between sham and tumor-bearing animals. The optimal running rate was identified by quantifying the amount of time spent running at or above certain rates using lower and upper specification limits generated by SAS Proc Capability. The optimal velocity was generated using SAS specification limits.

Measures of bone remodeling

Mice were X-rayed every week using a digital cabinet Faxitron Bioptics MultiFocus X-ray system. The X-ray images were provided to a blinded observer to perform longitudinal scoring of study endpoints using the following scales for assessing the extent of bone destruction (Eber et al, 2022): 0 = bones with no lesions; 1 = bones with one to three small pits of radiolucent lesions; 2 = bones with three to six small pits of radiolucent lesions; 3 = bones with obvious loss of medullary bone and erosion of cortical bone; 4 = bones with full-thickness unicortical bone loss; and 5 = bones with full-thickness bicortical bone loss and displaced skeletal fracture.

The trabecular bone was analyzed at the level of the distal femur and femoral neck, whereas the cortical bone was evaluated at distal femoral metaphysis using a micro-computed tomography (μCT) system (Skyscan 1272; Bruker). The scanning process was made at a 10 μm voxel size, and an X-ray power of 60 kVp and 166 μA with an integration time of 627 ms, according to the guidelines for μCT analysis of rodent bone structure (Bouxsein et al, 2010). Obtained images were reconstructed using NRecon software (Bruker). The trabecular ROI at distal femur metaphysis was evaluated by selecting 1 mm in the vertical axis, subsequent to 0.2 mm from the growth plate (reference point). For the cortical ROI analysis, the sample level was evaluated by selecting a band of 1 mm by 4 mm distal from the growth plate. For the femoral neck analysis, ROI was selected using a 0.5 mm² cylinder diameter, taking 1 mm of depth at 0.75 mm from the growth plate. The CT analyzer program (Bruker) was used to determine trabecular bone parameters; an automatic segmentation algorithm (CT analyzer) was applied to isolate the trabecular bone from the cortical bone. The parameters used for the trabecular bone were trabecular BMD, trabecular bone volume rate (BV/TV), trabecular thickness (Tb.Th), trabecular number (Tb.N), and trabecular separation (Tb.Sp). Cortical bone parameters in 3D were cortical BMD, cortical thickness (Ct.Th), and 2D cross-sectional cortical bone area (Ct.Ar). Finally, hydroxyapatite calibration phantoms (250 and 750 mg/cm³) were used to calibrate trabecular and cortical BMD values.

Subcutaneous injection mouse model

Luciferase-expressing DU145 human prostate cancer cells were inoculated subcutaneously into the right flank of immunodeficient mice. Tumor growth was measured via a caliper.

Animal tissue processing

At the termination of animal experiments, mice were perfused with 4% PFA. Spinal cords (L2-L4 region), DRGs (L2-L4), and femurs were collected, and these dissected tissues were prepared for immunohistochemistry and immunofluorescence (IF). In some cases, fresh serum, bone marrow supernatant, and DRGs were collected.

For frozen section preparation, spinal cords, DRGs, and femurs were post-fixed with 4% PFA for 12 h at 4°C, and then cryoprotected with 30% sucrose (Cat. #: S5-3; Thermo Fisher Scientific [Fisher Chemical]) solution in PBS (Cat. #: MT20031CV; Thermo Fisher Scientific [Gibco]) for 48 h at 4°C. Spinal cords and DRGs were embedded in Tissue-Plus optimum cutting temperature (Cat. #: 23-730-571; Thermo Fisher Scientific [Scigen Tissue-Plus]) compound, and frozen using dry ice. Femurs were decalcified with 10% EDTA (Cat. #: 327200025; Thermo Fisher Scientific [Fisher Chemical]) in pH 7.3 PBS for 14 d (the decalcification solution was replaced on day 8) before embedding in Tissue-Plus optimum cutting temperature compound and freezing. Frozen spinal cords, DRGs, and femurs were thaw-mounted on Superfrost non-adhesion glass slides (Cat. #: 12-550-143; Thermo Fisher Scientific) at 40 μm thickness, Superfrost Plus glass slides at 16 μm thickness, and Superfrost Plus glass slides (Cat. #: 12-550-16; Thermo Fisher Scientific) at 20 μm thickness, respectively.

For paraffin section preparation, the dissected femurs were post-fixed with 4% PFA for 6 h at 4°C, and then decalcified with 10% EDTA, in pH 7.3 PBS for 14 d (the decalcification solution was replaced on day 8) at 4°C. Thereafter, femurs were embedded in paraffin and then cut into 5-μm-thick sections.

In some experiments, fresh serum, bone marrow supernatant, and DRGs were collected without perfusion. To collect bone marrow supernatant, tibias and femurs were flushed with 500 μl PBS, the resultant fluid was centrifuged, and the clear supernatant was collected.

Nerve sprouting assay

Nerve sprouting assays were performed as previously described (Park et al, 2021). Briefly, a single-cell suspension of murine primary DRG sensory nerve cells was obtained from lumber DRG (L2-L4) of male C57BL/6 mice (8–12 wk of age) using enzymatic digestion and density gradient centrifugation. Then, 500–1,000 cells of DRGs in 30 μl of warm neuronal growth (NG) medium (Neurobasal-A [Cat. #: 21-103-049; Thermo Fisher Scientific (Gibco)], 1% N2 [Cat. #: 17502048; Thermo Fisher Scientific (Gibco)], 2% B-27 [Cat. #: 17-504-044; Thermo Fisher Scientific (Gibco)], 2 mM L-glutamine, 1% penicillin–streptomycin, and 0.4% glucose [MilliporeSigma]) were seeded onto the center of 14-mm round coverslips (Cat. #: P12G-1.5-14-F; MatTek Corp.), precoated with poly-D-lysine (50 μg/ml, overnight at 4°C, Cat. #: A3890401; Thermo Fisher Scientific [Gibco]) and laminin (20 μg/ml, 1 h at 37°C, Cat. #: CB-40232; Thermo Fisher Scientific [Corning]), in a 24-well plate. After 1–2 h, 1 ml of warm NG medium was gently added to the sides of wells and the cells were maintained at 37°C with 5% CO₂. After 48 h of primary DRG neuronal culture establishment, half (500 μl) of NG medium was replaced with 500 μl of either control or cancer cell–derived CM, and then, the DRG cells were incubated for another 48–72 h. In some cases, the DRG cells were treated with cancer CM, and either isotype or mouse anti-CGRP monoclonal antibody (100 nM; TEVA Pharmaceuticals). At the termination of the experiments, the DRG cells were fixed in 500 μl of 4% PFA (Cat. #: AA433689M; Thermo Fisher Scientific) for 10 min at room temperature, and immediately subjected to immunofluorescence or stored in 1X DPBS at 4°C until use. For each group, 2–3 coverslips were quantified, and 6–10 images were taken from each coverslip using a Nikon Eclipse Ni fluorescent microscope system (Nikon). Images were saved in nd2 or tiff files for further analysis using Visiopharm (Hørsholm) or ImageJ (NIH) software, respectively.

Immunofluorescence for CGRP, NF200, and tubulin β3 was performed on fixed DRG cells. After blocking with 5% normal donkey serum in 0.03% Triton X-100 in PBS for 1 h at room temperature, the cells were incubated with primary antibodies (anti-CGRP antibody [1:2,000, Cat. #: C8198; MilliporeSigma], anti-NF-200 antibody [1:3,000, Cat. #: CH22104; Neuromics], or anti-tubulin β3 [TUBB3] antibody [1:1,000, Cat. #: 801201; BioLegend]) overnight at 4°C. The cells were then labeled with the secondary antibodies (CY3-conjugated donkey anti-rabbit IgG [1:600, Cat. #: 711-165-152; Jackson ImmunoResearch], CY5-conjugated donkey anti-chicken IgY [1:400, Cat. #: 703-175-155; Jackson ImmunoResearch], or CY2-conjugated donkey anti-mouse IgY [1:500, Cat. #: 715-225-150; Jackson ImmunoResearch]) for 2 h at room temperature. After washing five times with 1X DPBS, the cells were mounted with ProLong Gold Antifade Mountant with DAPI. All images were taken using a Nikon Eclipse Ni fluorescent microscope system. For quantification, 6–10 images each coverslip x 2-3 coverslips were randomly selected, and the length of nerve fibers was analyzed, using image analysis software, Nikon Elements V4.13 Basic Research.

Co-culture between cancer cells and DRG cells

Cancer cells (2,000 cells/60 μl of growth medium supplemented with 10% FBS) were seeded onto the middle of μ-Slide 2 Well Co-Culture plates (Cat. #: 81806; ibidi). Murine primary DRG sensory nerve cells (1,000 cells/60 μl of NG medium) were then seeded onto the eight wells surrounding the tumor cells. After 24 h of incubation, 600 μl of serum-free media was added to allow these two cell types to exchange their secreted molecules. In some cases, co-cultures were treated with either isotype control (100 nM; TEVA Pharmaceuticals) or mouse anti-CGRP monoclonal antibody (100 nM; TEVA Pharmaceuticals). After 72 h of co-culture, cancer cells were stained with 1x crystal violet solution (Cat. #: S25275; Thermo Fisher Scientific [Fisher Science Education]) for 10 min at room temperature, rinsed with tap water five times, and lysed with 100 μl of 1% SDS solution (Cat. #: 11667289001; MilliporeSigma). The resulting cell lysates were transferred to a clear bottom 96-well plate (Cat. #: 07-200-565; Thermo Fisher Scientific [Corning]), and OD was measured at 570 nm by Epoch Microplate Spectrophotometer (Agilent Technologies). At the termination of the experiments, the resulting DRG cells were also fixed with 4% PFA for 10 min at room temperature, and the nerve growth was confirmed using IF.

Human samples

Serum samples were obtained from prostate cancer patients with/without bone metastasis from 2016 to 2018 (patients without [n = 11] and with [n = 22] bone metastasis), and ELISAs were used to measure levels of CGRP (Cat. #: CEA876Hu; Cloud-Clone Corp) and substance P (Cat. #: CEA393Hu; Cloud-Clone Corp) in plasma and bone marrow supernatant.

Human prostate adenocarcinoma tissue microarray (TMA) slides were sectioned (3 μm thickness) at the Jikei University (Tokyo, Japan). Samples were obtained from 93 patients who received prostatectomy in 2006. Greater than 90% of biopsy cores occupied by cancer were included for further analyses.

Autopsy bone samples obtained from prostate and lung cancer patients with/without bone metastases were collected during 2011–2015 and were sectioned at the Jikei University School of Medicine (5 μm thickness) (prostate cancer patients without [n = 4] and with [n = 4] bone metastasis and lung cancer patients without [n = 2] and with [n = 2] bone metastasis). Before incubating with the primary antibody, slides from patient samples were baked for 90 min at 60°C and treated with DeCal Retrieval solution (HK089-5K; BioGenex) for antigen retrieval for 30 min at room temperature.

After patient samples were blocked with 3% normal goat serum (Cat. #: 005-000-121; Jackson ImmunoResearch) and 0.3% Triton X-100 (Cat. #: X100-500ml; MilliporeSigma) in 0.01 M PBS for 1 h at room temperature, the TMA slides and bone slides were incubated with anti-CRLR primary antibody (1:50, Cat. #: 84467; Abcam) overnight at 4°C. Slides were then incubated with biotinylated anti-rabbit IgG secondary antibody (SS Rabbit Link, Cat. #: HK336-5R; BioGenex) for 1 h at room temperature. Antibody detection was performed using a VECTASTAIN ABC kit (Cat. #: PK-6100; Vector Laboratories). The CRLR-positive area (DAB intensity) was visualized using a Nikon Eclipse Ni fluorescent microscope. The CRLR-positive area was measured in three different randomly selected fields. The resulting DAB intensities were quantified by ImageJ software (Schindelin et al, 2012) by converting the DAB intensity number to an optical density (OD, OD = log [255 (max intensity)]/mean intensity).

Animal tissue immunohistochemistry and IF

Murine bone (paraffin): hematoxylin and eosin staining was performed on paraffin-embedded sections of femurs. IF for total p38 (1:200, Cat. #: 9212; Cell Signaling Technology) and phosphorylated p38 (1:200, Cat. #: 4511; Cell Signaling Technology) was performed on paraffin-embedded sections of femurs. Antigen retrieval was conducted using BioGenex DeCal Retrieval Solution (Cat. #: HK089-5K). After blocking with 3% normal donkey serum and 0.3% Triton X-100 in PBS for 1 h at room temperature, the sections were incubated with the first primary antibody, phosphorylated p38, overnight at 4°C. The sections were then labeled with the first secondary antibody CY5-conjugated donkey anti-rabbit IgG (1:600, Cat. #: 711-175-152; Jackson ImmunoResearch) at room temperature for 2 h. Sections were then blocked with 3% normal rabbit serum (Cat. #: 011-000-120; Jackson ImmunoResearch) and 0.3% Triton X-100 in PBS at room temperature for 2 h to saturate open primary antibody binding sites. Sections were blocked again with AffiniPure Fab fragment donkey anti-rabbit IgG (1:500, Cat. #: 711-007-003; Jackson ImmunoResearch) at room temperature for 2 h. Sections were then labeled with the second primary antibody, anti-total p38, overnight at 4°C. Finally, sections were labeled with the second secondary antibody, CY3-conjugated donkey anti-rabbit IgG (1:600), at room temperature for 2 h. Slides were then mounted for imaging. Other bones were stained for TRAP for osteoclasts (Cat#: 387A; Sigma-Aldrich). Slides were mounted using the aqueous mounting media, Aqua-Mount (Thermo Fisher Scientific), and slide edges were set with clear nail polish. Cells that were pinkish/red in color and along the proximal trabecular bone or proximal periosteum were counted as positive for TRAP as quantification for osteoclasts.

Murine bone (frozen): CGRP and NF200 (1:700, Cat. #: 4680; Abcam) staining was performed on frozen sections of the femur. After blocking with 3% normal donkey serum and 0.3% Triton X-100 in PBS for 1 h at room temperature, the frozen sections were incubated with primary antibodies (anti-CGRP and anti-NF200) overnight at 4°C. The sections were then labeled with the secondary antibodies CY3-conjugated donkey anti-rabbit IgG or CY2-conjugated donkey anti-chicken IgY (1:600, Cat. #: 703-225-155; Jackson ImmunoResearch) for 2 h at room temperature. Femur sections were mounted in ProLong Gold Antifade Mountant with DAPI (Cat. #: P36935; Thermo Fisher Scientific [Invitrogen]).

Murine spinal cord: IF for GFAP (1:2,000, Cat. #: Z0334; Dako), GFP (1:1,000, Cat. #: A-11122; Thermo Fisher Scientific [Invitrogen]), and CGRP was performed on fixed sections of the spinal cord. After blocking with 3% normal donkey serum and 0.3% Triton X-100 in PBS for 1 h at room temperature, the frozen sections were incubated with primary antibodies (anti-GFAP, anti-GFP, and anti-CGRP) overnight at 4°C. The sections were then labeled with the secondary antibodies CY3-conjugated donkey anti-rabbit IgG or CY2-conjugated donkey anti-chicken IgY for 2 h at room temperature. Spinal cord sections were dehydrated using different gradients of ethanol (70, 90, and 100%, for 2 min each), cleared with xylene (twice for 2 min each), and coverslipped with DPX mounting media (Cat. #: 06522; MilliporeSigma).

Murine DRG: fixed DRG sections were stained for GFP and CGRP. After blocking with 3% normal donkey serum and 0.3% Triton X-100 in PBS for 1 h at room temperature, the frozen sections were incubated with primary antibodies (anti-GFP and anti-CGRP) overnight at 4°C. The sections were then labeled with the secondary antibodies CY3-conjugated donkey anti-rabbit IgG or CY2-conjugated donkey anti-chicken IgY for 2 h at room temperature. DRG sections were mounted in ProLong Gold Antifade Mountant with DAPI.

All images were taken using a Nikon Eclipse Ni fluorescent microscope system (Nikon). Images were then quantified using Visiopharm (Hørsholm), Nikon Elements V4.13 Basic Research, and/or ImageJ software (version 1.51f; National Institutes of Health, Bethesda, MD).

Bioinformatics

The Cancer Genome Atlas was used to access genomics data available for prostate cancer cohorts (n = 390). Means and standard deviations were presented for continuous characteristics. Medians were also presented in the case that the continuous characteristics are not normally distributed. Counts and percentages were presented for discrete characteristics. The associations between the Gleason score and CALCRL, RAMP1, RAMP2, and RAMP3 were calculated using Spearman’s rank correlation coefficients using SAS software (SAS Inc., Cary, NC). The Gleason score was further categorized into three groups (i.e., 6, 7, and 8–10) based on clinically meaningfully cutoff points. Multinomial logistic regression was used to explore the association between CALCRL and the categorized Gleason score. Three models were performed. Model 1 was unadjusted. Model 2 adjusted for RAMP1 and age at diagnosis. Model 3 in addition adjusted for PSA. Odds ratios and their 95% confidence intervals were presented where the Gleason score 6 was treated as the reference group. Recurrence-free survival curves by CALCRL groups (using the median as a cutoff point) were computed using the Kaplan–Meier estimates and compared using log-rank tests. The association between the binary CALCRL (>median versus ≤median) and recurrence-free survival was performed using the Cox proportional hazards model. Three models were performed again with the same parameters. Hazard ratios for the association between binary CALCRL and recurrence-free survival and their 95% confidence intervals were calculated.

The GEO was used to access genomics data available for prostate cancer (GSE6919) (Yu et al, 2004; Chandran et al, 2007) and breast cancer (GSE14017 and GSE14018) (Zhang et al, 2009) cohorts. Log₂-transformed gene expression data of CALCRL, RAMP1, RAMP2, and RAMP3 from patient samples were compared between primary prostate cancer (n = 65) and metastatic prostate cancer (n = 25) obtained from GSE6919; between breast cancer with brain metastasis (n = 15), lung metastasis (n = 4), and bone metastasis (n = 10) obtained from GSE14017; and between breast cancer with brain metastasis (n = 7), lung metastasis (n = 16), liver metastasis (n = 5), and bone metastasis (n = 8) obtained from GSE14017 using GraphPad Prism 9 software (GraphPad Software Inc.).

Benign prostatic hyperplasia (n = 24), primary prostate cancer (n = 33), castration-resistant prostate cancer (CRPC) with soft tissue metastases (n = 129), and CRPC with bone metastases (n = 20) were obtained through the University of Washington Prostate Cancer Donor Autopsy Program (Morrissey et al, 2013). Briefly, tissue was sectioned, and RNA was isolated and amplified as we performed previously (Kumar et al, 2016). Probe labeling was performed using a custom Agilent 44K microarray kit (Cat. #: G4413A) against a reference pool of common prostate cancer cell lines. The CRPC molecular profiling data have been deposited in GEO with the accession number GSE77930 (Kumar et al, 2016).

ELISA

The levels of CGRP in the serum and bone marrow supernatant collected with protease inhibitors were measured with a custom CGRP ELISA (TEVA Pharmaceuticals). Briefly, mouse anti-CGRP capture antibody (Bertin Bioreagent, Montigny-le-Bretonneux) was coated on a 96-well plate and incubated overnight at 4°C. The coated plate was washed with wash buffer (PBS with 0.05% Tween-20) and then blocked using the Sword Blocker SBL-501 reagent (Sword Diagnostics) for 1 h. After several washes, serum samples and a range of CGRP standards diluted in Sword Diluent SDI-802 were added to the coated plate with a human anti-CGRP detection antibody (Teva Pharmaceuticals) and incubated at room temperature for 2 h. After multiple washes, captured CGRP analytes were complexed to an HRP-conjugated mouse anti-human antibody (Southern Biotech) at room temperature for 1 h. After several washes, Sword detector reagents consisting of a substrate/peroxidase mixture were used according to the manufacturer’s instructions. Resonance Raman signals generated using Sword reagents were measured at an excitation/emission wavelength of 530 nm/730 nm using a BioTek Cytation 5 microplate reader.

Quantitative Real-Time PCR (qRT-PCR)

Cells or DRG tissues were lysed, and RNA was harvested with RNeasy Mini Kit (Cat. #: 74104; QIAGEN). The RNA concentrations were determined and subsequently normalized between samples before first-strand cDNA synthesis. First-strand cDNA was synthesized using 0.5 μg of total RNA using Invitrogen SuperScript II Reverse Transcriptase (Cat. #: 18064022; Thermo Fisher Scientific [Invitrogen]). Real-time qPCR (qRT-PCR) was performed using TaqMan Gene Expression Master Mix (Cat. #: 4369016; Applied Biosystems) and TaqMan Gene Expression Assays (Cat. #: 4331182; Applied Biosystems, Assay IDs: Mm99999915_g1 [Gapdh, mouse GAPDH]; Mm00801463_g1 [Calca, mouse CGRP]; Hs02786624_g1 [Gapdh, human GAPDH]; Hs00907738_m1 [CALCRL, human CALCRL]; Hs00195288_m1 [RAMP1, human RAMP1]; Hs00237194_m1 [RAMP2, human RAMP2]; and Hs00389131_m1 [RAMP3, human RAMP3]). qRT-PCR was run for 50 cycles (95°C for 15 s and 60°C for 1 min) after an initial single cycle of 50°C for 2 min and 95°C for 10 min using CFX96 Touch Real-Time PCR Detection System (Bio-Rad). Data are presented as relative gene expression using the delta–delta Ct method, with Gapdh used as the reference gene.

Western blotting

Cells were lysed with cOmplete Lysis-M reagent (Cat. #: 4719956001; Roche) supplemented with cocktails of protease inhibitor (Cat. #: 11836170001; Roche) and PhosSTOP (Cat. #: 04906837001; Roche). In some cases, cancer cell lines (PC-3, DU145, MDA MB-231, and A549, 1 × 10⁶ cells/well in a six-well plate [1 ml]) were serum-starved for 5 h and then treated with 10–100 mM CGRP (Cat. #: 015-02; Phoenix Pharmaceuticals) and/or 1–5 μM SB203580 (Cat. #: S1076; Selleck Chemicals) for 5, 15, 30, and 60 min before cell lysis. In some cases, cells were pretreated with the CGRP receptor antagonist (1 nM, CGRP 8-37, Cat. #: 1181; Tocris Bioscience). After 5 min of boiling at 95°C, protein extracts (10–40 μg of protein per lane) were loaded, separated on SDS–PAGE (4–20% Tris–glycine gradient gels, Cat. #: XP04202BOX; Thermo Fisher Scientific [Invitrogen]), and transferred to a PVDF membrane (0.2 μm, Cat. #: ISEQ00010; MilliporeSigma). After electrophoresis, proteins were transferred to a 0.2-μm PVDF membrane. After blocking with 5% nonfat dry milk (Cat. #: 1706404XTU; Bio-Rad) in TBS/Tween-20 buffer (Cat. #: 28360; Thermo Fisher Scientific) for 1 h, the membrane was incubated with the following primary antibodies at 4°C overnight: anti-CRLR antibody (1:300, Cat. #: 84467; Abcam); anti-RAMP1 antibody (1:500, Cat. #: sc-11379; Santa Cruz Biotechnology); anti-RAMP2 antibody (1:500, Cat. #: ab198276; Abcam); anti-RAMP3 antibody (1:500, Cat. #: ab78017; Abcam), anti-p38 antibody (1:1,000, Cat. #: 9212; Cell Signaling Technology); anti-phospho-p38 (Thr180/Tyr182) antibody (1:1,000, Cat. #: 9211; Cell Signaling Technology); anti-HSP27 antibody (1:1,000, Cat. #: 95357; Cell Signaling Technology); anti-phospho-HSP27 antibody (1:1,000, Cat. #: 9709S; Cell Signaling Technology); or anti-GAPDH antibody (1:1,000, Cat. #: 2118; Cell Signaling Technology). Thereafter, blots were incubated with anti-rabbit IgG HRP-conjugated secondary antibody (1:2,000, Cat. #: 7074S; Cell Signaling Technology) at room temperature for 1 h. Protein expression was detected with Pierce ECL Western Blotting Substrate (Cat. #: 32209; Thermo Fisher Scientific). The densitometry analysis of the Western blot was performed with ImageJ software (National Institutes of Health [NIH], Bethesda, MD).

Cell proliferation assays

Cells were seeded into a 96-well plate in 100 μl complete medium at a concentration of 2 × 10³ cells/well. After 24 h of serum starvation, cells were incubated with CGRP (0–1μM; MilliporeSigma) for 48 h. MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide, 10 μl of 5 mg/ml) was added into each well and incubated for 4 h at 37°C, 5% CO₂, and 100% humidity, and then, 100 μl of the solubilization solution was added into each well for overnight. The absorbance of solubilized formazan crystals at 560–650 nm was measured with a microplate spectrophotometer. In some cases, cells were pretreated with the CGRP receptor antagonist (1 nM, CGRP 8-37, Cat. #: 1181; Tocris Bioscience) or p38 inhibitor (5 μM, SB203580, Cat. #: S1076; Selleckchem) for 1 h before the CGRP treatment. In other cases, cells were treated with isotype control and anti-CGRP Ab.

Cells were seeded into a 96-well plate in 100 μl complete medium at a concentration of 2 × 10³ cells/well. After 24 h of serum starvation, cells were incubated with CGRP (0–1 μM, 90954-53-3; MilliporeSigma) for 72 h. Then, plates were placed in IncuCyte ZOOM System (Essen Bioscience). Phase-contrast pictures were captured every 6 h. Cell confluency (%) and doubling time (hours) of each picture were automatically measured over time, and data were normalized with initial cell confluency. In some cases, cells were pretreated with CGRP 8-37 (dose) for 1 h before the CGRP treatment.

Intracellular signaling array

Cells were seeded into a six-well plate at 1 × 10⁵ cells/well. After 24 h of serum starvation, cells were treated with either dH₂O or 10 nM CGRP for 0.5, 1, 8, or 24 h. Then, the activation of signaling molecules in the cell lysates (0.3 mg/ml) was determined using PathScan intracellular Signaling Array Kit (Cat. #:7323; Cell Signaling Technology), according to the manufacturer’s instruction. The resulting spot intensities were quantified by ImageJ software. The data were normalized by total protein measured using the Pierce BCA protein assay kit (Cat #: 23227; Thermo Fisher Scientific).

Statistical analyses

Numerical data are expressed as the mean ± SD or SEM. Statistical analysis was performed using GraphPad Prism and SAS statistical program with significance at P ≤ 0.05. Outcome measures were transformed to satisfy the conditional normality assumption as needed. An unpaired t test or one-way analysis of variance (ANOVA) with Tukey’s or Dunnett’s post hoc test was used to compare single measurements between groups. For outcome measures (e.g., log-transformed radiance and adjusted guarding time) collected repeatedly over time, mixed-effects models were used to compare mean differences between groups (e.g., DU145 and Sham groups) over time. Group, time, and group-by-time interaction were included in the model. Animals were treated as a random effect. Contrasts were calculated to compare mean differences at each time point.

Online supplemental material

Fig S1 shows CALCA, RAMP1, RAMP2, and RAMP3 expression in one human prostate cancer cohort (GSE6919) and two human breast cancer cohorts (GSE14017, GSE14018). Fig S2 shows the relative mRNA and protein expression of CALCRL, RAMP1, RAMP2, and RAMP3 in human metastatic cancer cell lines. Fig S3 shows MTT of CGRP treatment on RM-1 cancer cells, depicts representative images of the antibody-based cell pathway array data initially shown in Fig 8A, and validates downstream targets in RM-1. Fig S4 illustrates μCT scans of femurs from WT mice and CGRP heterozygous (CGRP Control) and homozygous (CGRP KO) mice. In addition, this figure quantifies the trabecular area, cortical area, and femoral neck area in these bones. Fig S5 demonstrates no change in tumor growth between isotype control and anti-CGRP Ab in A549-bearing mice.